JP2007303842A - Vehicle position estimation device, and map information preparation device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a vehicle position estimation device capable of detecting precisely an absolute position of one's own vehicle without storing preliminarily a positional data of a landmark. <P>SOLUTION: This vehicle position estimation device detects, a plurality of times, the absolute position of the own vehicle and a relative position of the landmark in the periphery using the absolute position of the own vehicle as a reference, estimates an absolute position of the landmark with the smallest error, based on the absolute position and the relative position, and estimates the absolute position of the own vehicle, based on the absolute position of the landmark and the relative position of the detected landmark. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a vehicle position estimation device and a map information creation device, and in particular, estimates an absolute position of a mark around the own vehicle, and derives an absolute position of the own vehicle by calculating backward from the estimated absolute position of the mark. The present invention relates to a vehicle position estimation device and a map information creation device that creates map information based on an absolute position of a mark estimated by the vehicle position estimation device.

Conventionally, an error of about several meters to several tens of meters is included in the absolute position of the host vehicle detected by a GPS (global positioning system) device. As a technique for correcting this error and detecting an absolute position with high accuracy, Patent Document 1 stores in advance position data of a fixed object such as a reflector or an illumination lamp existing around a road, and the GPS device performs absolute detection. The position is detected and a fixed object around the road is detected by the radar device, and the position data of the fixed object detected by the radar device is compared with the position data of the fixed object stored in advance, and detected by the GPS device. A technique for correcting the absolute position is disclosed.
JP-A-10-300493

  However, when the accurate absolute position of the host vehicle is obtained using the technique of Patent Document 1, it is necessary to store in advance position data of landmarks such as fixed objects existing around all target roads. There was a problem that.

  Thus, in order to measure the position of the mark which exists around all the roads made into object, great cost is required. Further, since the amount of data stored in the vehicle position estimation device for storing the position data is enormous, it is not realistic.

  The present invention has been made to solve the above-described problems, and there is a vehicle position estimation device that can accurately detect the absolute position of the host vehicle without storing the position data of the mark in advance, and there is a road around it. It is an object of the present invention to provide a map information creating apparatus that can efficiently and efficiently collect position data of landmarks such as fixed objects to create map information used for creating a map.

  In order to achieve the above object, the invention described in claim 1 is directed to an absolute position detecting means for detecting an absolute position of the own vehicle with a predetermined accuracy, and a surrounding mark based on the absolute position of the own vehicle. The relative position detection means for detecting the relative position, the absolute position of the own vehicle detected by the self vehicle absolute position detection means and the relative position detection means, and the relative position of the mark. Estimating means for estimating the absolute position of the subject vehicle based on the absolute position of the mark and the relative position of the mark detected by the mark relative position detecting means; Storage means for storing estimated absolute position information indicating the absolute position of the mark estimated by the means, wherein the estimating means corresponds to the estimated absolute position corresponding to the mark to be estimated. If the information is stored in the storage means, in consideration of the absolute position indicated by the estimated absolute position information for estimating the absolute position of the mark.

  In the first aspect of the invention, the absolute position of the own vehicle is detected with a predetermined accuracy by the own vehicle absolute position detecting means, and the relative position of the surrounding marks based on the absolute position of the own vehicle is detected by the mark relative position detecting means. The position is supposed to be detected.

  In the present invention, the estimation means estimates the absolute position of the mark with the smallest error based on the absolute position and the relative position detected multiple times by the absolute position detection means and the relative position detection means, respectively. The absolute position of the vehicle is estimated based on the absolute position of the mark and the relative position of the mark detected by the mark relative position detection means, and the absolute position information indicating the absolute position of the mark estimated by the estimation means is stored by the storage means Is memorized. The storage means includes a RAM, a semiconductor memory such as a flash memory, a portable memory such as a compact flash (registered trademark) and an xD picture card (registered trademark), a fixed storage device such as a hard disk, or a wireless or wired network. The external storage device provided in the server computer etc. which were connected via this is contained.

  According to the present invention, the estimation means includes the absolute position indicated by the estimated absolute position information when the estimated absolute position information corresponding to the mark to be estimated is stored in the storage means. Estimate the absolute position of the placemark.

  Thus, according to the first aspect of the present invention, the absolute position of the host vehicle and the relative position of the surrounding landmarks based on the absolute position of the host vehicle are detected a plurality of times, and based on the absolute position and the relative position. The absolute position of the mark with the smallest error is estimated, and the absolute position of the vehicle is estimated based on the absolute position of the mark and the relative position of the detected mark. Therefore, the position data of the mark is stored in advance. In addition, the absolute position of the host vehicle can be detected with high accuracy.

  According to the first aspect of the present invention, the estimated absolute position of the mark is stored in the storage means, and the absolute position of the mark is estimated by taking into account the absolute position of the mark indicated by the estimated absolute position information. Therefore, it is possible to suppress a decrease in accuracy of the absolute position of the mark estimated in the next and subsequent estimations.

  In the present invention, as in the invention described in claim 2, the estimation means further derives accuracy information indicating the estimation accuracy of the estimated absolute position of the mark, and the storage means derives from the estimation means. Only estimated absolute position information whose estimated accuracy indicated by the accuracy information is a predetermined accuracy or higher may be stored.

  According to a second aspect of the invention, as in the third aspect of the invention, the storage unit further stores the accuracy information in association with the estimated absolute position information, and the accuracy derived by the estimation unit. Update means for updating the estimated absolute position information stored in the storage means when the estimated accuracy indicated by the information is higher than the estimation precision indicated by the accuracy information stored in the storage means May be.

  According to a first aspect of the invention, as in the fourth aspect of the invention, the host vehicle absolute position detecting means detects the absolute position of the host vehicle in a state including a biased error that periodically changes. The storage means stores a plurality of estimated absolute position information estimated by the estimation means at different timings with respect to the same mark, and the estimation means stores a plurality of estimated absolute positions stored in the storage means. Based on the absolute position of each mark indicated by the position information, the absolute position of the mark with the smallest error is re-estimated and detected by the mark relative position detecting means on the basis of the re-estimated absolute position of the mark. The absolute position of the host vehicle may be estimated based on the relative position of the mark.

  Further, according to the present invention, the storage means may store the absolute position of the host vehicle estimated by the estimation means.

  In the present invention, as in the invention described in claim 6, the mark relative position detecting means may detect the relative position by at least one of a laser radar, a millimeter wave radar, and a camera.

  Further, according to the present invention, as in the invention described in claim 7, the mark is a pillar-shaped object, a corner or ridge portion of a building, a cut portion of a guard rail, a manhole on the road, a reflector provided on the guard rail, or embedded in a road surface. It may include at least one of a road ridge, a road marking, and a curb cut.

  On the other hand, the map information creating apparatus according to claim 8 is configured to detect a relative position of a surrounding marker with reference to the absolute position of the own vehicle, and an absolute position detecting unit for detecting the absolute position of the own vehicle with a predetermined accuracy. The relative position detection means for detecting, the absolute position of the own vehicle detected by the own vehicle absolute position detection means and the relative position detection means of the host vehicle and the relative position of the mark, respectively, with the smallest error. Estimating means for estimating the absolute position of a mark, and non-volatile storage means for storing estimated absolute position information indicating the absolute position of the mark estimated by the estimating means as map information for creating a map. ing.

  According to the eighth aspect of the present invention, the absolute position of the host vehicle is detected with a predetermined accuracy by the host vehicle absolute position detecting unit, and the relative position of the surrounding markers based on the absolute position of the host vehicle is detected by the marker relative position detecting unit. The position is supposed to be detected.

  In the present invention, the estimation unit has the smallest error based on the absolute position of the host vehicle and the relative position of the mark detected multiple times by the host vehicle absolute position detecting unit and the mark relative position detecting unit, respectively. The absolute position of the mark is estimated, and non-volatile storage means stores estimated absolute position information indicating the absolute position of the mark estimated by the estimating means as map information for creating a map. The nonvolatile storage means includes a nonvolatile semiconductor memory such as a flash memory, a portable memory such as a compact flash (registered trademark) and an xD picture card (registered trademark), a fixed storage device such as a hard disk, or a wireless device. And an external storage device provided in a server computer connected via a wired network.

  Thus, according to the eighth aspect of the invention, the absolute position of the own vehicle and the relative position of the surrounding landmarks based on the absolute position of the own vehicle are detected a plurality of times, and based on the absolute position and the relative position. The absolute position of the mark with the smallest error is estimated, and the estimated absolute position information indicating the estimated absolute position of the mark is stored in the non-volatile storage means, so the position of the mark such as a fixed object existing around the road Map information can be created by collecting data accurately and efficiently.

  As described above, according to the vehicle position estimation apparatus of the present invention, the absolute position of the host vehicle and the relative position of the surrounding landmarks based on the absolute position of the host vehicle are detected a plurality of times, and the absolute position and the relative position are detected. The absolute position of the mark with the smallest error is estimated based on the absolute position of the host vehicle based on the absolute position of the mark and the relative position of the detected mark. There is an excellent effect that the absolute position of the host vehicle can be detected with high accuracy without storing.

  Further, according to the map information creation device of the present invention, the absolute position of the host vehicle and the relative position of the surrounding landmarks based on the absolute position of the host vehicle are detected a plurality of times, and an error is determined based on the absolute position and the relative position. The absolute position of the smallest landmark is estimated, and the estimated absolute position information indicating the estimated absolute position of the landmark is stored in the non-volatile storage means, so the position data of the landmark such as a fixed object existing around the road It is possible to create map information used for creating a map by accurately and efficiently collecting the information.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

[First Embodiment]
FIG. 1 shows the configuration of a vehicle position estimation apparatus 10 according to the first embodiment.

  As shown in the figure, the vehicle position estimation device 10 includes a host vehicle absolute position detection unit 12 that detects the absolute position of the host vehicle with a predetermined accuracy, and a three-dimensional object around the host vehicle based on the absolute position of the host vehicle. A landmark relative position detection unit 14 for detecting the relative position of a landmark (hereinafter referred to as a landmark), a vehicle absolute position detection unit 12 and a landmark relative position detection unit 14 are connected to each other, and is configured by a microcomputer or the like. The apparatus 20 is provided.

  In the present embodiment, a general-purpose GPS device having an error of about several meters to several tens of meters is used as the host vehicle absolute position detection unit 12. This GPS device receives signals from a plurality of GPS satellites and detects the absolute position on the earth such as the latitude, longitude, and altitude of the host vehicle, and controls absolute position information indicating the detected absolute position. Output to the device 20. Note that a gyro sensor may be used as the host vehicle absolute position detection unit 12 instead of the GPS device, or a hybrid type using both the GPS device and the gyro sensor may be used. In the following, the latitude and longitude of the host vehicle detected by the GPS device are represented by a two-dimensional XZ coordinate system (X, Z).

  In the present embodiment, the landmark relative position detector 14 includes a light emitting element that emits an infrared light pulse and a light receiving element that receives an infrared light pulse reflected from a front landmark. A laser radar attached to the front grille or bumper of the vehicle is used. This laser radar calculates the time until the infrared light pulse reflected by the landmark is received by the light receiving element and the infrared light pulse based on the time when the light emitting element emits light and the infrared light pulse is irradiated. Based on the irradiated angle, the relative position of the landmark represented by the distance r to the landmark relative to the host vehicle and the azimuth angle θ of the landmark with respect to the traveling direction of the host vehicle is detected with an accuracy of several centimeters of error. can do. The landmark relative position detection unit 14 detects the relative positions of surrounding landmarks with reference to the host vehicle at a predetermined cycle (for example, 10 Hz), and controls the relative position information indicating the detected relative positions. Output to.

  The landmark relative position detection unit 14 may be a millimeter wave radar or a camera instead of the laser radar. Moreover, you may use what combined these arbitrarily. In the case of using a camera, the same area may be imaged a plurality of times while moving with one camera, and the distance r to the landmark may be obtained from the captured images using a moving stereo method. Further, it is possible to calculate the distance r according to the position of the landmark in the image based on the camera parameters, assuming that all the images captured by one camera are on a plane. . When the landmark in the image can be determined as a three-dimensional object from the features such as the shape, the distance r to the run mark may be obtained on the assumption that the lower end is on a plane. Furthermore, a plurality of cameras that capture the same region at a predetermined distance may be provided, and the distance r to the landmark may be obtained from a plurality of images captured by each camera by a stereo method.

  In this embodiment, as a landmark, a pole-like object such as a utility pole, a sign, a traffic signal, a corner of a building, a ridgeline part of a building, a break part of a guardrail, a manhole on a road, a reflector provided on a guardrail, and embedded in a road surface It is assumed that there is a road ridge (so-called cat's eye), road marking, curb cut, etc., but this is not a limitation, and any object that does not change the position around the road It may be.

  A columnar object is an object that appears relatively frequently around the road, and its thickness is within a certain range. There are advantages such as easy to obtain. For example, in a laser radar, a reflected signal as shown in FIG. 4 is obtained, and a columnar object is often obtained as an independent point having a relatively strong intensity. In laser radar and millimeter wave radar, columnar object candidates are extracted based on changes in reflection intensity and time series, grouping of reflection points, and the like. In addition, since a monochrome image is obtained with a monochrome camera, and a gray image including color information of R (red), G (green), and B (blue) is obtained with a color camera, the density value for each pixel of the obtained image is obtained. When edges are detected by differentiation or the like, and two vertical edges are detected within a predetermined width, it can be detected as a columnar object.

  The corner of the building refers to the edge portion formed by the surface of the building, and the ridgeline of the building refers to the outline that forms the boundary with the background. When a building such as a residential area is close to a road or in a group of buildings such as a downtown area, corners and ridges of buildings around the road are effective detection targets. The corners of buildings are easy to detect with a laser radar or the like, and the edges are easy to detect with a camera. The millimeter wave radar is difficult to detect the corners and ridgelines of a building, but when a metal object is installed in the building, a reflected wave from the metal object can be obtained.

  The pole part of the guardrail is difficult to correlate in chronological order because objects of the same shape are continuously arranged, but in the case of a break part of the guardrail, it indicates the end of the continuous columnar object, and it is easy to specify the position, It can be detected by millimeter wave radar, laser radar and camera.

  Although the appearance frequency of manholes on the road is not so high, it is easy to identify because it has a circular or square shape, and can be detected by a millimeter wave radar, a laser radar, and a camera.

  Reflectors provided on guardrails and road fences buried on the road surface are easily reflected even at night and can be detected by a laser radar and a camera. Further, when a metal part is attached to a reflector or a roadside, it can be detected by a millimeter wave radar.

  Road markings (for example, intersection marks, temporary stop lines, diamond marks, rhombus marks indicating the presence of pedestrian crossings, maximum speed markings, arrows indicating the direction of travel according to the direction of travel, pedestrian crossings, etc.) Even if the number of lanes is large, it exists for each lane, so that it is easy to use it as a landmark even when it is difficult to detect buildings and columnar objects around the road. The camera can identify the road marking by determining the shape based on the captured image. In laser radar, a road marking can be detected by irradiating a laser toward the road to obtain a reflection signal from a white painted portion of the road marking.

  When the road marking is detected, for example, the road marking can be identified by extracting the feature point that is the apex of the road marking from the differential value of the density value for each pixel and determining the arrangement of the feature points.

  When detecting lane markings (white lines, etc.) as road markings, for example, white line breaks, white lines in areas that can be overtaken, white lines in compound lines that are visual guide lines, white lines in branches and merges, white line joints, What is necessary is just to detect the start point end point of a zebra zone, etc.

  The edge of the curb can be detected with a laser radar and a camera.

  That is, when a laser radar or a millimeter wave radar is used as the landmark relative position detection unit 14, an infrared light pulse or millimeter reflected by the object is utilized by utilizing a difference in reflectance due to a difference in shape or material of the object. The object can be detected by identifying the wave intensity. In the case of laser radar, columnar objects, building corners and ridges, guardrail cuts, manholes on the road, reflectors on guardrails, roads Can detect traps, road markings, curb cuts, etc.In the case of millimeter wave radar, columnar objects, building corners and ridges, guardrail cuts, manholes on the road, reflectors provided on the guardrail, Road traps can be detected.

  In addition, when a camera is used as the landmark relative position detection unit 14, an image processing such as filtering or differentiation is performed on the captured image to detect an edge and identify the shape of the object. (For example, when two vertical edges are detected within a predetermined width in a captured image, it is detected as a columnar object.) A columnar object, a corner or ridge portion of a building, It is possible to detect a break in the guardrail, a manhole on the road, a reflector provided on the guardrail, a road fence, a road marking, a curb break, and the like.

  FIG. 2 is a block diagram illustrating a functional configuration of the control device 20 according to the present embodiment.

  As shown in the figure, the control device 20 includes a host vehicle position evaluation unit 22 for deriving information related to movement such as the speed of the host vehicle and detection accuracy of the absolute position, and a landmark such as the type of the detected landmark. A relative position evaluation unit 24 for deriving information on the attributes of the vehicle and a relative position detection accuracy; a temporary storage unit 26 for temporarily storing various types of information; and the own vehicle absolute position detection unit 12 and the landmark relative position detection unit 14. The absolute position of the landmark with the smallest error is estimated based on the absolute position of the own vehicle and the relative position of the landmark each detected a plurality of times, and is detected by the absolute position of the landmark and the landmark relative position detector 14. A position estimation unit 32 that estimates the absolute position of the host vehicle based on the relative position of the landmark to be detected, and information such as the estimated absolute position of the landmark. It includes a non-volatile memory device 40 such as an HDD for storing and the.

  The host vehicle position evaluation unit 22 performs control to store a predetermined number of pieces of absolute position information input in the temporary storage unit 26 in the newest order. When the absolute position information is input, the absolute position indicated by the input absolute position information. Is stored in the temporary storage unit 26 as detected vehicle position information.

  In addition, the host vehicle position evaluation unit 22 stores in advance a standard variance value of the absolute position detected by the host vehicle absolute position detection unit 12, and when the absolute position information is input, the standard variance stored in advance is stored. The value is stored in the temporary storage unit 26 as absolute position evaluation information indicating the absolute position detection accuracy. Further, when the absolute position information is input, the host vehicle position evaluating unit 22 outputs a signal indicating that the absolute position information has been input to the position estimating unit 32.

  In the GPS device, the absolute position detection accuracy may change depending on the positioning conditions. For this reason, for example, the map information is stored in the storage device 40 in advance, and the vehicle position evaluation unit 22 obtains the position indicated by the absolute position information based on the map information, and the position is similar to that of the countryside. If it is a place, the dispersion value σa may be changed, and if it is a place like an urban center where a building stands, the dispersion value σb (> σa) and the detection accuracy may be changed. In addition, when a value (for example, DOP (Dilution of Precision)) indicating detection accuracy is input together with absolute position information from the GPS device, the host vehicle position evaluation unit 22 may derive a variance value based on the value. . Further, as shown in FIG. 3, the host vehicle position evaluation unit 22 obtains a least square approximation straight line of each absolute position indicated by a predetermined number of absolute position information stored in the temporary storage unit 26, and uses the input absolute position information. It is good also as detection accuracy by calculating | requiring the difference of the absolute position shown and a least squares approximate line.

  The relative position evaluation unit 24 stores in advance an expression for obtaining a detection accuracy corresponding to the distance r and the azimuth angle θ detected by the landmark relative position detection unit 14, and when the relative position information is input, The relative position of the landmark indicated by the relative position information is stored in the temporary storage unit 26 as detected relative position information, and the relative position based on the detection accuracy corresponding to the relative position (distance r and azimuth angle θ) of the landmark. And the variance value is stored in the temporary storage unit 26 as relative position evaluation information. In addition, when the relative position information is input, the relative position evaluation unit 24 outputs a signal indicating that the relative position information has been input to the position estimation unit 32. When the landmark relative position detection unit 14 is a camera, the detection accuracy differs depending on the distance between the landmark and the camera. Therefore, parameters such as the number of pixels of the camera and the focal length are stored in advance and based on them. Deriving the variance of the detected relative position. The relative position evaluation unit 24 may evaluate reliability based on the intensity of the signal detected by the landmark relative position detection unit 14 and add reliability information to the variance of the detected relative positions. Good.

  The position estimation unit 32 uses an extended Kalman filter (explained later) to detect detected vehicle position information, detected relative position information stored in the temporary storage unit 26, and the absolute position of the landmark estimated last time (described later). The absolute position of the landmark with the smallest error based on the accuracy of the estimated absolute position information of the landmark, the absolute position of the own vehicle estimated last time, and the absolute position accuracy of the own vehicle estimated last time, A variance value indicating the accuracy of the absolute position, an absolute position of the host vehicle, and a variance value indicating the accuracy of the absolute position of the host vehicle are estimated. Then, the position estimation unit 32 uses the estimated absolute position of the landmark as landmark estimated absolute position information, the variance of the absolute position of the landmark as landmark estimated position evaluation information, and determines the estimated absolute position of the host vehicle. The vehicle absolute position information is used, and the variance value of the absolute position of the host vehicle is stored in the temporary storage unit 26 as host vehicle position evaluation information.

  In addition, the position estimation unit 32 stores the absolute position of the landmark indicated by the landmark estimated absolute position information and the accuracy of the absolute position in the storage device 40 as landmark information according to the estimated absolute position accuracy of the landmark. The absolute position of the host vehicle indicated by the host vehicle absolute position information is controlled according to the accuracy of the absolute position of the host vehicle indicated by the host vehicle position evaluation information stored in the temporary storage unit 26. Control which memorize | stores in the memory | storage device 40 as traveling locus information is performed.

  Next, an extended Kalman filter that simultaneously estimates the absolute position and variance of the landmark and the absolute position and variance of the host vehicle employed in the position estimation unit 32 according to the present embodiment will be described.

  In the present embodiment, using the extended Kalman filter represented by the following relational expressions (1) to (5), the absolute position of the landmark and the absolute position of the host vehicle estimated one time ago are detected (observed). Based on the absolute position of the subject vehicle, the relative position of the landmark, and the respective variances, the absolute position of the subject vehicle and the landmark that are estimated to have the smallest error variance at the present time are estimated.

  It is assumed that the estimated value regarding the estimated position of the host vehicle and the landmark has a parameter of the following equation (6), and the detected observation value has a parameter of the following equation (7).

  If the state transition F of the movement of the host vehicle is defined as in the following equation (8), the relationship between the detected observed value and the estimated value is expressed as in the following equation (9). Since the extended Kalman filter is used in the present embodiment, the observation matrix is defined as the equation (10) linearized around the predicted value.

Here, for example, when only one landmark is detected (i = 1), x = (X _V0 , Z _V0 , φ _V , V _V , ω _V , X _L1 , Z _L1 ) is estimated. It becomes a parameter, and the error covariance matrix of the estimated value at this time is shown as the following equation (11).

_Four error covariance matrices σ ² _XL ¹ _{XL 1} , σ ² _XL ¹ _{ZL 1} , σ ² _ZL ¹ _{XL 1} , and σ ² _ZL ¹ _{ZL 1} represent the dispersion of the landmark.

  Note that it is necessary to perform an operation with an initial value determined, and it is necessary to carefully set a variance value when a landmark is detected for the first time. The variance has the relationship of the following expression (12). The variance of the (m + 1) th landmark when m landmarks have been detected and the (m + 1) th landmark is detected next is (13).

Next, the operation of the vehicle position estimation device 10 for deriving the absolute position of the host vehicle will be described. The host vehicle absolute position detection unit 12 receives signals from a plurality of GPS satellites and receives the signals on the earth of the host vehicle. The absolute position is detected, and absolute position information indicating the absolute position is output to the control device 20 every time the absolute position is detected.

  On the other hand, the landmark relative position detection unit 14 detects the relative positions of surrounding landmarks with reference to the host vehicle in a predetermined cycle, and the relative position information indicating the relative position is detected every time the relative position is detected. 20 output.

  FIG. 4 is a plan view in which the relative positions of the landmarks indicated by the relative position information detected by the landmark relative position detector 14 are plotted.

  When the absolute position information is input from the own vehicle absolute position detection unit 12, the own vehicle position evaluation unit 22 determines the input absolute position information and a standard variance value of the absolute position stored in advance as absolute position evaluation information in a new order. Only a single signal is stored in the temporary storage unit 26, and a signal indicating that absolute position information has been input is output to the position estimation unit 32.

  When the relative position information is input from the landmark relative position detection unit 14, the relative position evaluation unit 24 stores the input relative position information in the temporary storage unit 26 as detected relative position information and also stores the standard of the relative position stored in advance. The relative position variance is stored in the temporary storage unit 26 as relative position evaluation information, and a signal indicating that the relative position information has been input is output to the position estimation unit 32.

  The position estimation unit 32 is connected to an external device (not shown). When a signal designating the start of detection processing is input from the external device, the position estimation unit 32 is based on a predetermined number of absolute position information stored in the temporary storage unit 26. After deriving an initial value of the current speed V of the host vehicle, a position estimation process described later is started.

  FIG. 5 occupies a flowchart showing the flow of position estimation processing executed in the position estimation unit 32. This position estimation process is executed when a signal designating the start of the detection process is input to the position estimation unit 32 from an external device (not shown).

  In step 100 of the figure, it is determined whether or not a signal indicating that the relative position information is input from the relative position evaluation unit 24 is input. If the determination is affirmative, the process proceeds to step 102 and the determination is negative. If YES, go to step 110.

  In the next step 102, it is determined whether or not a signal indicating that absolute position information has been input from the host vehicle position evaluation unit 22 has been input. If the determination is affirmative, the process proceeds to step 106 and a negative determination is made. If YES in step 104, the flow advances to step 104.

  In step 104, since the absolute position of the host vehicle has not been detected, the absolute position used for estimation is predicted from the above-described formula (3) based on the host vehicle position evaluation information estimated one time (Δt) before, Using the predicted absolute position of the host vehicle, the process of step 106 described later is performed on the assumption that the observed value of the absolute position of the host vehicle cannot be obtained.

  In the next step 106, using the above-described extended Kalman filter, the detected host vehicle position information, the detected relative position information, the landmark estimated absolute position information estimated one time ago, the landmark, stored in the temporary storage unit 26, Based on the estimated position evaluation information, the own vehicle absolute position information, and the own vehicle position evaluation information, the absolute position of the landmark that is estimated to have the smallest error, the variance value of the absolute position of the landmark, and the absolute position of the own vehicle , And the variance value of the absolute position of the host vehicle.

  That is, the absolute position of the host vehicle detected by the host vehicle absolute position detection unit 12 and the relative position with respect to the landmark based on the host vehicle detected by the landmark relative position detection unit 14 include errors. However, by repeating the processing from step 100 of the host vehicle position estimation processing to step 122 described later to repeat estimation of the absolute position of the landmark, estimation is performed as shown in <1> to <3> of FIG. Since the error (variance) in the absolute position of the landmark to be reduced is reduced, the accuracy of the absolute position of the host vehicle estimated based on the absolute position and relative position information of the landmark is also improved.

  Note that the position estimation unit 32 searches whether the landmark information corresponding to the landmark is stored in the storage device 40 when estimating the absolute position of the landmark detected using the extended Kalman filter for the first time. If it is stored, the absolute position indicated by the landmark information and the accuracy of the absolute position are taken into account as an initial value, and the estimation by the extended Kalman filter is performed. If not stored, the above equation (12), (13 ) Estimation using the extended Kalman filter is performed using the landmark position variance calculated by the equation (1) as an initial value.

  In step 106, the estimated absolute position of the landmark is used as landmark estimated absolute position information, the variance of the absolute position of the landmark is used as landmark estimated position evaluation information, and the estimated absolute position of the own vehicle is used as the absolute position of the own vehicle. As the position information, the variance value of the absolute position of the host vehicle is stored in the temporary storage unit 26 as the host vehicle position evaluation information.

  In the next step 108, the following landmark information storage process is performed to control the storage device 40 to store information relating to the landmark.

  FIG. 7 is a flowchart showing the flow of landmark information storage processing.

  In step 150 in the figure, it is determined whether or not the landmark information corresponding to the landmark has already been registered in the storage device 40. If the determination is affirmative, the process proceeds to step 156, and the determination is negative. If YES in step 152, the flow advances to step 152.

  In step 152, it is determined whether or not the variance σ indicated by the landmark estimated position evaluation information is less than the predetermined value ThL. If the determination is affirmative, the process proceeds to step 154. The landmark information storage process ends, and the process returns to the host vehicle position estimation process. That is, when the accuracy of the absolute position of the estimated landmark is high, the variance σ indicated by the landmark estimated absolute position information is small. Therefore, in step 152, the accuracy of the absolute position of the landmark estimated by determining whether the variance σ is less than the predetermined value ThL is evaluated. Thereby, for example, when a moving object such as a pedestrian is detected as a landmark by the laser radar, the dispersion σ becomes large, so that the landmark information can be prevented from being registered in the storage device 40. The predetermined value ThL may be set in advance to an appropriate value in accordance with the required accuracy of the absolute position of the landmark.

  In step 154, the absolute position of the landmark indicated by the landmark estimated absolute position information stored in the temporary storage unit 26 and the variance value of the absolute position indicated by the landmark estimated position evaluation information are stored as landmark information. 40 and return to the vehicle position estimation process.

  On the other hand, in step 156, it is determined whether or not the variance σ indicated by the landmark estimated position evaluation information is less than the variance σrL indicated by the landmark information already stored in the storage device 40. In this case, the process proceeds to step 158. If the determination is negative, the landmark information storage process ends, and the process returns to the host vehicle position estimation process.

  In the next step 158, the absolute position of the landmark indicated in the landmark information stored in the storage device 40 and the variance of the absolute position are indicated in the landmark estimated absolute position information stored in the temporary storage unit 26. The updated absolute position of the landmark and the variance of the absolute position indicated by the estimated landmark position evaluation information are returned to the vehicle position estimation process.

  That is, when the accuracy of the estimated landmark estimated absolute position information is higher than the accuracy of the landmark information stored in the storage device 40, the landmark information stored in the storage device 40 is updated.

  On the other hand, in step 110 of the own vehicle position estimation process (see FIG. 5), as in step 102 described above, it is determined whether or not a signal indicating that absolute position information has been input from the own vehicle position evaluation unit 22 has been input. If the determination is affirmative, the process proceeds to step 112. If the determination is negative, the process proceeds to step 114.

  In step 112, since the relative position of the landmark has not been detected, the absolute position used for estimation is estimated based on the own vehicle position evaluation information estimated one time ago (Δt) and the detected current own vehicle absolute position information. Using the extended Kalman filter excluding parameters related to landmark observation information, the absolute position of the host vehicle that is estimated to have the smallest error and the variance of the absolute position of the host vehicle are estimated, and the estimated absolute position of the host vehicle is calculated. As the own vehicle absolute position information, the variance value of the absolute position of the own vehicle is stored in the temporary storage unit 26 as the own vehicle position evaluation information.

  On the other hand, in step 114, since the absolute position of the host vehicle and the relative position of the landmark are not detected, the current position is calculated from the above-described formula (3) based on the host vehicle position evaluation information estimated one time (Δt) ago. The absolute position of the host vehicle is predicted, and the predicted absolute position of the host vehicle is overwritten and stored in the temporary storage unit 26 as host vehicle absolute position information. Further, based on the variance value indicated by the own vehicle position evaluation information, the value calculated by the equation (5) is set as the predicted variance value.

  In the next step 120, the host vehicle absolute position information and host vehicle position evaluation information stored in the temporary storage unit 26 are read, and the host vehicle absolute position information and host vehicle position evaluation information are output to an external device.

  In the next step 121, the following running track storage process is performed to control the storage device 40 to store the running track of the host vehicle.

  FIG. 8 occupies a flowchart showing the flow of the travel locus storage process.

  In step 170 of the figure, it is determined whether or not the variance σ indicated by the own vehicle position evaluation information is less than the predetermined value ThC. If the determination is affirmative, the process proceeds to step 172, and the determination is negative. Ends the travel locus storage process and returns to the vehicle position estimation process. That is, in this step 170, the accuracy of the estimated absolute position of the host vehicle is evaluated. The predetermined value ThC may be set in advance to an appropriate value according to the required absolute position accuracy of the host vehicle.

  In step 172, the absolute position of the host vehicle indicated by the host vehicle absolute position information stored in the temporary storage unit 26 is stored in the storage device 40 as travel locus information, and the process returns to the host vehicle position estimation process.

  That is, when the accuracy of the estimated absolute position of the host vehicle is high, the absolute position is stored in the storage device 40. Thereby, it is possible to record an accurate travel locus of the host vehicle.

  In step 122 of the host vehicle position estimation process (see FIG. 5), it is determined whether or not a signal for designating the end of the detection process has been input to the position estimation unit 32 from an external device (not shown). If the determination is negative and the determination is negative, the process waits for a predetermined time Δt from the current time t and returns to step 100 when the current time reaches t + Δt.

  As described above, the processing of Step 100 to Step 122 according to the present embodiment is repeated, and the estimation of the absolute position of the landmark using the above-described extended Kalman filter is performed a plurality of times for the same landmark. The accuracy of the absolute position of the estimated landmark is improved, and the accuracy of the absolute position of the host vehicle is also improved.

  Further, when the landmark information has already been stored in the storage device 40 by the previous travel of the host vehicle, the absolute position of the landmark indicated by the landmark information by the extended Kalman filter and the variance of the absolute position are used as initial values. Thus, the absolute position of the landmark can be accurately estimated from the beginning.

  FIG. 9 is a schematic diagram showing changes in the dispersion of the absolute positions of the landmarks estimated by the subject vehicle position estimation process.

  In the first run, the error of the absolute position of the landmark estimated by the extended Kalman filter is relatively large due to the error due to the GPS device or the error of the laser laser, and the absolute position of the host vehicle estimated from the absolute position of the landmark is The error is relatively large.

  However, after the second run, the error of the absolute position of the landmark is reduced by performing the estimation by the extended Kalman filter with the landmark information estimated by the vehicle position estimation processing until the previous time as an initial value. Accordingly, the error of the absolute position of the host vehicle estimated from the absolute position of the landmark can also be reduced.

  Further, according to the present embodiment, it is possible to create a highly accurate map showing landmark positions around the road using the landmark information stored in the storage device 40.

  Furthermore, the number of lanes of the road can be determined by referring to the travel locus on the road indicated by the travel locus information stored in the storage device 40 after traveling the same road a plurality of times. It is also possible to create a road map with better accuracy based on landmark information and travel locus information.

  As described above, according to the first embodiment, the absolute position of the host vehicle is detected with a predetermined accuracy by the host vehicle absolute position detecting means (here, the host vehicle absolute position detecting unit 12), and the mark relative position is detected. The relative position of surrounding landmarks based on the absolute position of the host vehicle is detected by the detection means (here, the landmark relative position detection unit 14).

  And according to 1st Embodiment, the absolute position of the own vehicle detected by the estimation means (here, the position estimation part 32) several times each by the own vehicle absolute position detection means and the mark relative position detection means and The absolute position of the mark with the smallest error is estimated based on the relative position of the mark, and the absolute position of the host vehicle is estimated based on the absolute position of the mark and the relative position of the mark detected by the mark relative position detecting means. Therefore, the absolute position of the host vehicle can be detected with high accuracy without storing the mark position data in advance.

  According to the first embodiment, the storage means (here, the storage device 40) stores the estimated absolute position information indicating the absolute position of the mark estimated by the estimation means, and the estimation means When the estimated absolute position information corresponding to the mark is stored in the storage means, the absolute position of the mark is estimated by taking into account the absolute position indicated by the estimated absolute position information. A decrease in the accuracy of the absolute position of the mark estimated at the time of estimation can be suppressed.

  Further, according to the first embodiment, the estimation unit further derives accuracy information indicating the estimation accuracy of the estimated absolute position of the mark, and the storage unit is indicated by the accuracy information derived by the estimation unit. Since only the estimated absolute position information with the estimated accuracy equal to or higher than the predetermined accuracy is stored, it is possible to prevent the estimated absolute position information of the mark from being stored when a moving object is detected as the mark.

  Further, according to the first embodiment, the storage unit further stores the accuracy information in association with the estimated absolute position information, and the estimation accuracy indicated by the accuracy information derived by the estimation unit is stored in the storage unit. In the case where the accuracy is higher than the estimation accuracy indicated by the accuracy information, the update unit (here, the position estimation unit 32) that updates the estimated absolute position information stored in the storage unit is further provided. The accuracy of the estimated absolute position information of the mark is improved by estimating the absolute position of the mark.

  In the present embodiment, the case where both the absolute position of the landmark and the absolute position of the host vehicle are estimated using the extended Kalman filter has been described. However, the present invention is not limited to this, for example, a particle filter. May be adopted. Further, for example, the own vehicle absolute position detection unit 12 and the landmark relative position detection unit 14 detect the absolute position of the own vehicle and the relative position of the landmark with reference to the own vehicle a plurality of times, and the detected own vehicle The absolute position of the landmark may be estimated by obtaining the absolute position of the landmark a plurality of times based on the absolute position and the relative position of the landmark, and obtaining an average value of the obtained absolute positions of the plurality of landmarks. Further, the position where the sum of the Euclidean distances of the absolute positions of the plurality of landmarks is minimized may be estimated as the absolute position of the landmark. The absolute position of the host vehicle may be estimated based on the relative position of the landmark based on the host vehicle detected by the landmark relative position detection unit 14 with the estimated absolute position of the landmark as a reference. Good. Regarding the absolute position variance, the absolute position variance of a plurality of landmarks estimated in the past may be used, or the average value of the absolute position variance of a plurality of landmarks estimated in the past may be taken.

[Second Embodiment]
Here, the absolute position detected by the GPS device may include a deviation due to the positioning of the GPS satellite (orbital deviation), the ionosphere and the troposphere delay, etc., within a limited time, There is a case where a deviation occurs in the average of a plurality of detected absolute positions.

  Therefore, in the second embodiment, when the GPS device detects the absolute position of the host vehicle including a biased error, the embodiment corrects the bias of the error and accurately detects the position of the host vehicle. Will be described.

  Since the schematic configuration of the vehicle position estimation apparatus 10 according to the second embodiment is the same as that of FIG. 1, the description thereof is omitted.

  FIG. 10 shows a functional configuration of the control device 20 according to the second embodiment. Note that the same components in FIG. 2 as those in FIG. 2 are denoted by the same reference numerals as those in FIG.

  The position estimation unit 32 according to the second embodiment does not update the landmark information in the landmark information storage process described above, and indicates the landmark estimated absolute position information estimated every time the host vehicle travels. The absolute position of the landmark and the variance of the absolute position indicated by the landmark estimated position evaluation information are added to the storage device 40 and stored as landmark information.

  Then, the position estimation unit 32 re-estimates the absolute position of the landmark based on the absolute position of the landmark indicated by each landmark information stored in the storage device 40 and the accuracy of the absolute position, and detects the error most. The absolute position of the landmark that is estimated to be small and the variance of the absolute position are obtained and stored in the storage device 40 as landmark reference information. The absolute position of the landmark may be obtained by averaging the absolute position of the landmark indicated by each landmark information.

  As a result, even if an error is included in the absolute position of each landmark estimated due to the influence of an error having a periodic bias in the absolute position of the host vehicle detected by the GPS device, it is estimated at different timings. By re-estimating the absolute position of the landmark from the absolute position of the landmark, the influence of the periodic error can be reduced.

  The position estimation unit 32 uses the absolute position of the landmark indicated by the landmark reference information re-estimated and stored in the storage device 40 as a reference position, and the relative position of the vehicle relative to the landmark based on the detected relative position information. Is used to estimate the absolute position of the host vehicle. This improves the accuracy of the absolute position of the host vehicle even when the GPS device detects a biased error.

  As described above, according to the second embodiment, the own vehicle absolute position detection means detects the absolute position of the own vehicle in a state including a biased error that periodically changes, and the storage means , Storing a plurality of estimated absolute position information estimated by the estimating means at different timings for the same mark, and the estimating means stores the absolute values of the marks respectively indicated by the plurality of estimated absolute position information stored in the storage means Based on the position, the absolute position of the mark having the smallest error is re-estimated, and the self-position is detected based on the relative position of the mark detected by the mark relative position detection means with the re-estimated absolute position of the mark as a reference. Since the absolute position of the vehicle is estimated, the absolute position of the host vehicle is accurately detected even when the absolute position detected by the absolute position detecting means includes a biased error that periodically changes. It can be.

  The configurations of the vehicle position estimation device 10 and the control device 20 (see FIGS. 1, 2, and 7) described in the first and second embodiments are examples, and do not depart from the gist of the present invention. Needless to say, it can be appropriately changed within the range.

  Further, the flow of the vehicle position estimation process (see FIG. 5) described in the present embodiment is also an example, and it is needless to say that it can be changed as appropriate without departing from the gist of the present invention.

It is a block diagram which shows schematic structure of the vehicle position estimation apparatus which concerns on 1st Embodiment. It is a block diagram which shows the function structure of the control apparatus which concerns on embodiment. It is a graph which shows an example of the least square approximate straight line of an absolute position. It is a figure which shows an example of the position of the landmark detected by the laser radar. It is a flowchart which shows the flow of the own vehicle position estimation process which concerns on embodiment. It is the figure which showed typically the flow of the own vehicle position estimation. It is a flowchart which shows the flow of the landmark information storage process which concerns on embodiment. It is a flowchart which shows the flow of the driving | running | working locus | trajectory memory | storage process which concerns on embodiment. It is the figure which showed typically the change of dispersion | distribution of the absolute position of the estimated landmark at the time of drive | working in multiple times. It is a block diagram which shows the function structure of the control apparatus which concerns on 2nd Embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Vehicle position estimation apparatus 12 Own vehicle absolute position detection part 14 Landmark relative position detection part 26 Temporary storage part 32 Position estimation part 40 Storage device 50 Map information storage part

Claims (8)

Own vehicle absolute position detecting means for detecting the absolute position of the own vehicle with a predetermined accuracy;
Mark relative position detection means for detecting the relative position of surrounding marks based on the absolute position of the vehicle;
Estimating the absolute position of the mark with the smallest error based on the absolute position of the own vehicle and the relative position of the mark detected multiple times by the own vehicle absolute position detecting means and the mark relative position detecting means, respectively. Estimating means for estimating the absolute position of the host vehicle based on the absolute position of the mark and the relative position of the mark detected by the mark relative position detecting means;
Storage means for storing estimated absolute position information indicating the absolute position of the mark estimated by the estimating means;
With
When the estimated absolute position information corresponding to the mark to be estimated is stored in the storage means, the estimating means determines the absolute position of the mark in consideration of the absolute position indicated by the estimated absolute position information. A vehicle position estimation device for estimation. The estimation means further derives accuracy information indicating the estimation accuracy of the estimated absolute position of the landmark,
The vehicle position estimation apparatus according to claim 1, wherein the storage unit stores only estimated absolute position information whose estimation accuracy indicated by the accuracy information derived by the estimation unit is equal to or greater than a predetermined accuracy. The storage means further stores the accuracy information in association with the estimated absolute position information,
The estimated absolute position stored in the storage means when the estimated accuracy indicated by the accuracy information derived by the estimation means is higher than the estimation accuracy indicated by the accuracy information stored in the storage means The vehicle position estimation apparatus according to claim 2, further comprising update means for updating information. The host vehicle absolute position detecting means detects the absolute position of the host vehicle in a state including a biased error that periodically changes,
The storage means stores a plurality of estimated absolute position information estimated by the estimation means at different timings for the same mark,
The estimation means re-estimates the absolute position of the mark with the smallest error based on the absolute position of the mark respectively indicated by the plurality of estimated absolute position information stored in the storage means, and re-estimates the mark The vehicle position estimation apparatus according to claim 1, wherein the absolute position of the host vehicle is estimated based on a relative position of the mark detected by the mark relative position detection means with reference to the absolute position of the mark. The vehicle position estimation apparatus according to any one of claims 1 to 4, wherein the storage means stores an absolute position of the host vehicle estimated by the estimation means. The vehicle position estimation device according to any one of claims 1 to 5, wherein the mark relative position detection means detects the relative position by at least one of a laser radar, a millimeter wave radar, and a camera. The landmark is at least one of a pillar-shaped object, a corner or ridge of a building, a break of a guard rail, a manhole on the road, a reflector provided on the guard rail, a road fence buried in the road surface, a road marking, and a curb cut. The vehicle position estimation apparatus according to any one of claims 1 to 6. Own vehicle absolute position detecting means for detecting the absolute position of the own vehicle with a predetermined accuracy;
Mark relative position detection means for detecting the relative position of surrounding marks based on the absolute position of the vehicle;
Estimating means for estimating the absolute position of the mark having the smallest error based on the absolute position of the own vehicle and the relative position of the mark detected respectively by the own vehicle absolute position detecting means and the mark relative position detecting means. When,
Nonvolatile storage means for storing estimated absolute position information indicating the absolute position of the mark estimated by the estimating means as map information for creating a map;
Map information creation device with

Images